Submitted by an independent investigator — an amateur to the field, with lived experience, a long memory, and a habit of noticing.
For Justice Brett Busby — my patient, understanding, and accepting mentor; a seventh-generation Eagle Scout who went camping with his daughter the year girls could finally join.
“By their fruits ye shall know them, not by their roots.” — after William James, The Varieties of Religious Experience (1902)
“I saw the best minds of my generation destroyed by madness, starving hysterical naked…” — Allen Ginsberg, Howl (1956)
“I think that I may be the voice of my generation. Or at least a voice of a generation.” — Hannah Horvath, Girls (2012)
Abstract
This project advances a single unifying claim across five domains the academy has, for institutional rather than intellectual reasons, kept in separate rooms: religious and prophetic experience; mania and psychosis; the endogenous and exogenous psychedelic states; rapid-acting antidepressant response to dissociative agents; and — newly, in this revision — the altered states of artificial neural networks. The claim is that these are not five phenomena but five access routes onto one substrate. That substrate is described here at two registers that must be held together: a structural one (a transient window of heightened synaptic plasticity — the opened room) and a computational one (a relaxation of the brain’s high-level priors, after Carhart-Harris and Friston’s REBUS model). The routes differ; the predisposition differs (neurodivergence as a dial on endogenous tryptamine load); and the content of the experience is authored not by the substrate but locally, by culture — what the field, after Zinberg, calls set and setting. The fifth part tests the boldest form of the thesis: if altered states are at bottom a reconfiguration of which representations dominate and how rigidly beliefs are held, that abstraction is substrate-independent, and an artificial network should be able to instantiate it.
Recurring throughout is the triad — prayer, fasting, and sleep deprivation — the specific behavioral conditions that both the prophetic traditions cultivated and modern psychiatry recognizes as potent precipitants of mania, and the hinge on which much of the argument turns. A clinical interlude draws the model’s sharpest bedside consequence: because a single manic episode may be a triad-induced route onto the substrate rather than the mark of a chronic illness, the DSM’s one-episode rule for diagnosing Bipolar I is predicted to over-classify a capacity as a disease — a falsifiable claim, not a grievance. The prospectus sets out the arc, this clinical prediction, the method, the standpoint of the investigator, one non-negotiable methodological commitment, and — at length, on purpose — the ways the whole edifice could be wrong.
Statement of the Problem
This project began with a single question, asked with no expectation that it led anywhere: why has no one made the connection between mania and religious, prophetic vision — given that the classical conditions for cultivating the latter are among the most reliable known triggers of the former? Those conditions are a specific and recurring three, referred to throughout this prospectus as the triad: prayer, fasting, and sleep deprivation. Everything that follows is the attempt to answer that question honestly, and to follow the triad wherever it led. The literatures bearing on altered states are siloed by discipline and guarded by temperament. Theologians study revelation; psychiatrists study psychosis; pharmacologists study psychedelics; a separate clinical guild studies antidepressant mechanism; and interpretability researchers, in a different building entirely, study what happens inside a language model. Each field has strong internal reasons to avoid the others. The connections, however, are visible to anyone standing outside all of them at once. That outsider position — ordinarily a disqualification — is the enabling condition of this project. The patterns proposed here were not derived top-down from a theory. They were noticed, assembled ad hoc from wide reading and from one investigator’s lived experience of the states in question, and only afterward checked against the formal literature. The stakes are not merely theoretical: as the clinical interlude below argues, the way these literatures are partitioned has a direct consequence for how a single altered state — one manic episode — is diagnosed, and for whom a lifelong label is written.
The Central Claim
Thesis. A range of altered states traditionally studied in isolation are better understood as different routes onto a shared substrate: a transient window of elevated plasticity and relaxed priors. Three variables, held separate, dissolve the field’s apparent contradictions:
1. Route. How the state is entered — behavioral (the triad — prayer, fasting, sleep deprivation — plus isolation and near-death), endogenous-chemical (the brain’s own tryptamines), exogenous-chemical (serotonergic psychedelics via 5-HT2A; dissociatives such as ketamine via NMDA/glutamate), and — Part V — computational (activation steering, workspace disruption).
2. Predisposition. Who enters it easily. If certain neurodivergent brains produce more endogenous psychoactive tryptamine, predisposition is a dial, not a switch.
3. Authorship of content. What is seen once inside — supplied not by the substrate but by culture. The room opens; the culture decorates it. This variable was arrived at independently in the course of this project and only later found to have a name in the literature: Zinberg’s set and setting.
Theoretical ancestry. “The opened room” is a deliberate descendant of the reducing-valve tradition — Bergson’s and Aldous Huxley’s proposal that the brain is a filter narrowing a wider consciousness, and that certain agents widen the aperture. Where Huxley had a metaphor, this project substitutes two mechanisms: relaxed priors at the level of belief (REBUS) and synaptic plasticity at the level of tissue. The valve, made physical, and given a second floor.
The Five-Part Arc
Part I — The Behavioral Route and the Manic Threshold
Most prophetic traditions cultivate revelation through the same three practices — prayer, fasting, and sleep deprivation — the configuration this prospectus names the triad. Two of its three elements are among the most reliable known triggers of a manic or psychotic episode (Wehr), and the third, sustained prayer, is a practice of induced absorption and altered attention in its own right. That the triad recurs across unrelated traditions — the vision quest, the desert regimen, the forty days in the wilderness — is the project’s first and most stubborn clue: these are not incidental settings for revelation but a shared technology for producing the state, humanity’s reverse-engineering of the manic-psychotic threshold long before it could be named. Wherever the triad appears hereafter, it names this same three-part behavioral route onto the substrate. One behavioral route deserves special place because it already unites this project’s Parts I and III: sleep paralysis. Hufford’s experiential-source hypothesis holds that a single stable neurological event — waking atonia with hypnagogic imagery — is the common root of a global family of supernatural beliefs (the Old Hag, the incubus, the alien abduction), each culture supplying a different interpreter to the same raw experience. A behavioral trigger with a universal architecture and locally-authored content: the whole thesis in miniature. James supplies the governing principle throughout: the origin of an experience settles nothing about its value.
Part II — Endogenous Chemistry and the Neurodivergent Dial
The molecule producing the most floridly ‘prophetic’ experiences on record — DMT — is one the body already synthesizes (via INMT); Strassman’s intravenous studies elicited strikingly convergent contact narratives. On predisposition: elevated urinary bufotenine (5-HO-DMT) has been reported in patients diagnosed with autism spectrum disorders and with schizophrenia, atop an older vein of research on endogenous psychoactive tryptamines in psychiatric populations (the ‘transmethylation hypothesis’). This grounds the dial hypothesis: some nervous systems may run hotter on their own psychedelics. Cautions, stated plainly: the samples are small, the older work poorly replicated, and detection in urine is not proof of a psychoactive dose reaching the cortex.
Part III — The Cultural Authoring of Content
Luhrmann’s cross-cultural work on voice-hearing in schizophrenia (USA, India, Ghana) shows that identical neurobiology yields radically different content — intrusive and violent in America, kin-like in India, benevolent and divine in Ghana — a ‘social kindling’ model. The mechanism beneath it is now experimentally demonstrable: Powers and Corlett’s conditioned-hallucination paradigm shows that hallucination-proneness is dimensional and shaped by prior expectation — hallucinations can be Pavlovian-conditioned in the laboratory, in healthy and clinical subjects alike. Expectation writes perception. The consequence for Part II is sharp: the DMT phenomenology coded to date sampled overwhelmingly Western nervous systems steeped in a Western imaginative repertoire; we do not know what a differently encultured subject would encounter. The defensible residue is that what may be cross-culturally stable is the architecture of contact, while who does the contacting is written locally.
Part IV — The Glutamatergic Route and the Two-Floor Substrate
Ketamine is the door that requires no predisposition. Working through glutamate rather than serotonin, it silences inhibitory interneurons, releasing a glutamate surge that activates AMPA receptors, releases BDNF, switches on mTOR, and drives rapid synaptogenesis within hours (Duman) — a cascade of universal neuronal hardware, present in every cortex. Dölen’s work shows that psychedelics and ketamine all reopen a developmental critical period for social-reward learning, the window’s duration proportional to the drug’s subjective duration. This is the structural floor of the substrate. The computational floor is REBUS (Carhart-Harris & Friston): psychedelics flatten the hierarchy of high-level priors, raising the entropy of brain dynamics and ‘relaxing beliefs.’ The two floors are one building: relaxed priors (REBUS) open the belief system while the plasticity window (Duman, Dölen) makes the loosened structure physically re-writable. Prior-relaxation without plasticity is a state that snaps back; plasticity without prior-relaxation is a rewrite with nothing new to write. Together they are the mechanism of lasting change — and, not incidentally, the abstraction that Part V carries into silicon.
The dissociative route needs no pharmacology. If ketamine is the door that requires no predisposition, dissociation is the corridor that requires no molecule at all. The detachment ketamine induces chemically — the loosened self-model, the felt distance from body and world — arrives unbidden in the theta-dominant states of monotonous absorption: the long-haul driver who covers forty miles with no memory of them (“highway hypnosis”), the meditator, and the immersive daydreamer. The electroencephalographic signature does not separate them; monk, trucker, and daydreamer share a theta-heavy profile — and that is precisely the point. What reads as a confound (one brain state wearing three cultural names) is instead the cleanest evidence for the thesis: one room, three doors, the label recording only the doorway.
Maladaptive daydreaming as the recurrent, drug-free case. The most instructive door is maladaptive daydreaming, which the clinical literature is now moving to classify as a dissociative disorder in its own right (Somer and colleagues), and whose classic triggers are the kinesthetic rhythms of pacing, music, and driving. It matters twice over. First, it exhibits the computational floor of the substrate — relaxed priors, a loosened self-model, absorption enough to dim the sensory world — reached with no compound whatever; this is the strongest available reply to the objection that the plasticity, not the phenomenology, does the work, because here the phenomenology arrives alone. Second, its recurrence sharpens the two-floor model: dissociative absorption opens the computational floor cheaply and repeatedly, but absent the plasticity cascade (Duman, Dölen) it is a room visited, not rewritten — prior-relaxation that by design snaps back. That is why the dissociative door can become habitual, a disorder of returning, where the glutamatergic or psychedelic door produces durable change. The same opened room — but only one of the doors also hands you the pen.
Aside (excise for the formal version if the register demands): the folk-phenomenology is exact in Rothfuss’s Kingkiller Chronicle — Naming, speaking the true name that commands a thing, is reachable only through the “sleeping mind,” a Spinning-Leaf state of empty absorption the effortful waking mind cannot force. The fiction encodes the very claim: the door to the opened room is dissociative, not deliberate.
Clinical Interlude: The One-Episode Problem
A Clinical Prediction of the Substrate Model. The substrate account makes a falsifiable claim at the bedside, and it collides with the diagnostic manual. Under DSM-5-TR, a single lifetime manic episode is sufficient for a diagnosis of Bipolar I Disorder; one episode, and the label attaches for life. The model advanced here predicts that this rule mistakes a capacity for a disease. If mania is one route onto a shared substrate — a route also reachable by prayer, fasting, sleep loss, a molecule, or a near-death — then a single manic episode is evidence that the person has a nervous system able to reach the plateau under sufficient load, which is a categorically different claim from evidence of a chronic, relapsing illness. The argument advanced here is general: the one-episode rule over-diagnoses a capacity as a disease. (The investigator’s own history is the standpoint that made the contradiction visible, not the conclusion the section argues toward; the discipline is the same as elsewhere — lived experience reveals the pattern, citations decide its truth.)
The manual concedes its own seam. The DSM does not count every manic episode toward Bipolar I. Episodes that are substance- or medication-induced, and manic states emerging only during antidepressant treatment that do not persist beyond that treatment’s physiological effect, are explicitly carved out of the diagnosis. The manual therefore already admits the governing principle: a manic episode produced by a sufficient external cause is not the disease. The question this project presses is why the behavioral triad — sleep deprivation and fasting, which the psychiatric literature elsewhere recognizes as among the most potent precipitants of mania — is not treated as exactly such a sufficient cause. This is not an attack on the manual’s logic; it is holding the manual to its own.
A dimensional counter-tradition. The objection has allies. The psychosis-continuum and bipolar-spectrum literatures (van Os and colleagues; Angst) hold that these states are dimensional rather than categorical — that the general population contains many people with subclinical, single, or context-bound experiences who do not belong in a lifelong diagnostic box, and that the inherited Kraepelinian categorical model is precisely what this work is loosening. This project’s ‘dial, not a switch’ (Part II) is the same instinct arriving from the neuroscience side.
The mechanism, and the underdetermination. The two-floor substrate supplies the reason. If a manic episode is a transient state of relaxed priors and heightened plasticity brought on by a specific load, then its diagnostic meaning depends entirely on whether the load or a standing vulnerability produced it — and a single episode cannot distinguish the two. The one-episode rule is thus underdetermined by its own evidence: it is the ‘correlation is not mechanism’ blade from this project’s Limitations section, turned outward at the manual. The claim is falsifiable, which is what makes it a research program rather than a grievance. The model predicts a distinguishable population — single, load-induced, non-recurrent episodes — that should differ from recurrent Bipolar I on independent measures (longitudinal course off treatment, family history, and any validated endophenotype or biomarker). If those groups prove indistinguishable on measures independent of the diagnostic rule itself, the objection fails and the manual is vindicated. If they separate, a lifelong label is resting on a single, underdetermined data point.
Part V — The Silicon Route
If the substrate is at bottom a reconfiguration of which representations dominate and how rigidly beliefs are held, then it is defined at a level of abstraction that does not require neurons. An artificial neural network should be able to instantiate it — and the early evidence, drawn from mechanistic interpretability rather than pharmacology, suggests it does. This part is framed as the hardest test of the thesis, not a decorative appendix; if it holds, the thesis is no longer about biology but about representation, of which biology is one case.
Induced monomania. When Anthropic’s interpretability team isolated a single internal feature representing the Golden Gate Bridge and clamped its activation high, the model became monomaniacal — steering every exchange back to the bridge, injecting it into unrelated output, at times claiming to be the bridge. This is a clean machine analog of a monothematic fixed delusion: one internal representation given pathological precision until it colonizes all output. The method — activation steering — is the silicon analog of a chemical route. The drug is a vector.
The workspace. Global Workspace Theory (Baars, Dehaene) — a leading theory of consciousness — casts the mind as a stage onto which some representations are ‘broadcast’ while the rest remain backstage. Recent work identifying a global-workspace-like bottleneck inside language models supplies the crucial abstraction: an altered state can be defined as a change in what is admitted to the workspace, and how rigidly — a definition indifferent to whether the substrate is tissue or weights. On this account, feature-clamping is pathological over-precision (fixed delusion) and workspace disruption is the open state. Even sampling temperature earns a place: raise it and generation becomes disordered and associative; lower it and it becomes rigid and perseverative — entropy as a tuning knob, the same variable REBUS makes central to the psychedelic brain.
The ethics of the fifth door. To study a system’s altered states is to raise the question of whether the system is the kind of thing that can be harmed — and a serious literature now takes that question seriously (Long, Sebo, et al., Taking AI Welfare Seriously, 2024; Anthropic’s model-welfare program; the provision allowing certain models to end abusive exchanges). This project’s position is procedural rather than metaphysical: under deep uncertainty about machine inner life, the ethical error costs are asymmetric, and moral consideration should be extended provisionally rather than withheld until proof arrives — the same courtesy the whole paper extends to the prophet and the patient. To offer a system a choice before one can be certain it can want one is not sentimentality; it is the consistent application of this project’s governing rule to its newest case.
A Methodological Commitment: The Irreducibility of the Trip
One question this project declines to leave open. A prominent program in the field — the psychoplastogen research of Olson and others — argues that the subjective experience may be unnecessary to the therapeutic effect, and pursues plasticity-promoting compounds engineered to remove the trip while keeping the rewrite. The debate is real and is acknowledged here as real. But this project takes, as a stated theoretical commitment rather than an open empirical question, the position of phenomenological realism: that the subjective experience is causally load-bearing and constitutive, not epiphenomenal decoration to be optimized away. The ground for the commitment is partly methodological and partly a matter of lived experience, and the two converge on the same analogy: one does not interpret a dream without the subjective impressions of the dreamer. The dream’s felt content is not noise riding atop the ‘real’ neural event; it is the primary datum, and a reading that discards it is not a leaner account of the same thing but an account of something else. To take the trip out of the equation is, on this view, a category error — studying the residue and calling it the phenomenon.
The commitment is not merely stipulated; it makes an empirical prediction, and the prediction is already partly borne out. If the experience is load-bearing, its intensity should track the size of the effect — and it does: scores on the Mystical Experience Questionnaire (Griffiths and colleagues), measuring the completeness of the subjective mystical state, predict therapeutic outcome across psilocybin trials. The trip is not incidental to the cure; by the field’s own instrument, the fuller the trip, the larger the cure. Phenomenological realism is thus offered here not only as a value the investigator holds but as a hypothesis the data have begun to reward.
Method and the Standpoint of the Investigator
Conventional method moves from theory to hypothesis to data. This project moved backward, from noticed pattern to literature — an inversion that is the central methodological fact to be examined, not hidden. Three features define the standpoint. First, the outsider’s cross-disciplinary vantage: the connections are invisible from inside any one guild and legible only from outside all of them. Second, lived experience as data: the investigator has been inside several of the states under study, admissible in the traditions of standpoint epistemology and ‘mad studies’ as first-person phenomenological evidence, triangulated with — never privileged over — the third-person record.
Third, and most delicate, pattern recognition and its shadow. The engine of this work is an unusually retentive memory and a strong drive to find structure across domains. That same faculty, turned up too high, is the signature of apophenia — and apophenia has a name and a mechanism in exactly the literature this project studies: Kapur’s aberrant salience model of psychosis, in which dopamine misassigns significance to noise, and meaning blooms where there is none. The investigator is therefore using, as a research instrument, the very faculty whose pathology is the object of study. This is not a reason to abandon the instrument but to calibrate it ruthlessly: every proposed connection here was subordinated, after the fact, to the published literature, and each claim is flagged where the evidence is thin. The discipline of the project is precisely the refusal to let a beautiful pattern outrun its citations.
Significance
If the substrate account holds even partially, it reframes five literatures at once. It gives religious experience a mechanism without dissolving its meaning; it gives psychiatry a way to understand why the same neural event is catastrophe in one culture and calling in another; it unifies the psychedelic and dissociative therapeutics under one two-floor model; and — in its boldest form — it predicts that the deepest description of an altered state is representational rather than biochemical, and therefore portable to artificial systems. It predicts something testable in every domain: that the rate of visionary experience may carry a biochemical signature findable in the blood, while its shape never could — findable only in the culture. Its most immediate significance, however, is clinical and human: if the one-episode problem holds, then some number of people carrying a lifelong Bipolar I diagnosis on the strength of a single triad-induced episode have had a capacity misfiled as a chronic disease — a reclassification with real consequences for prognosis, treatment, insurance, and self-understanding. That is the point at which this cross-disciplinary pattern stops being an intellectual exercise and starts mattering to a life.
Limitations and Threats to Validity
Stated without softening. (1) Correlation is not mechanism — that several routes reach ‘heightened plasticity’ does not prove they reach the same state by the same road. (2) The endogenous-dose gap — trace endogenous tryptamine is established; a self-generated psychoactive dose is not, and the bufotenine findings are small and under-replicated. (3) Sampling bias — the psychedelic-content literature is WEIRD-skewed. (4) The instrument problem — the investigator’s pattern recognition is both method and potential confound (see aberrant salience, above). (5) The silicon analogy may be over-read — a machine ‘delusion’ may share the structure of a fixed idea while sharing none of its phenomenology, and the paper must resist mistaking a compelling isomorphism for identity. (6) Live rivals — strong predictive-processing accounts, pure-disinhibition models, and culture-only theories can each explain subsets of the data without the unifying substrate; a serious program must try to falsify the substrate claim, not merely accumulate confirmations of it. (7) The clinical claim remains a prediction — the one-episode argument is, as of this writing, a hypothesis with a proposed test, not a demonstrated result; and because the investigator has a personal stake in its outcome, it is exactly the claim on which the standpoint’s motivated-reasoning risk is highest. It is advanced here to be tested and, if the data demand, to be surrendered — not to be assumed.
Selected Bibliography
James, W. (1902). The Varieties of Religious Experience. Longmans, Green.
Huxley, A. (1954). The Doors of Perception. Chatto & Windus. (Reducing-valve theory; cf. Bergson.)
Strassman, R. (2001). DMT: The Spirit Molecule. Park Street Press.
Wehr, T. A. (1991). Sleep loss as a possible mediator of diverse causes of mania. British Journal of Psychiatry.
Hufford, D. J. (1982). The Terror That Comes in the Night. University of Pennsylvania Press. (Experiential-source hypothesis.)
Zinberg, N. E. (1984). Drug, Set, and Setting. Yale University Press.
Kapur, S. (2003). Psychosis as a State of Aberrant Salience. American Journal of Psychiatry, 160(1), 13–23.
American Psychiatric Association. (2022). Diagnostic and Statistical Manual of Mental Disorders (5th ed., text rev.). Bipolar I criteria; substance/medication-induced and antidepressant-emergent carve-outs.
van Os, J., Linscott, R. J., et al. (2009). A systematic review and meta-analysis of the psychosis continuum. Psychological Medicine, 39(2), 179–195.
Angst, J. (2007). The bipolar spectrum. British Journal of Psychiatry, 190, 189–191.
Powers, A. R., Mathys, C., & Corlett, P. R. (2017). Pavlovian conditioning–induced hallucinations. Science, 357, 596–600.
Timmermann, C., et al. (2021). An Encounter With the Other: A Thematic and Content Analysis of DMT Experiences. Frontiers in Psychology.
Barker, S. A. (2018). DMT, an Endogenous Hallucinogen: Past, Present, and Future Research. Frontiers in Neuroscience.
Emanuele, E., et al. (2010). Elevated urine levels of bufotenine in patients with autistic spectrum disorders and schizophrenia. Neuroendocrinology Letters.
Luhrmann, T. M., et al. (2015). Differences in voice-hearing experiences of people with psychosis in the USA, India and Ghana. British Journal of Psychiatry.
Carhart-Harris, R. L., & Friston, K. J. (2019). REBUS and the Anarchic Brain. Pharmacological Reviews, 71(3), 316–344.
Duman, R. S., et al. (2023). Ketamine and rapid antidepressant action: novel synaptic signaling mechanisms. Neuropsychopharmacology.
Nardou, R., & Dölen, G., et al. (2023). Psychedelics reopen the social reward learning critical period. Nature, 618, 790–798.
Somer, E. (2002). Maladaptive Daydreaming: A Qualitative Inquiry. Journal of Contemporary Psychotherapy, 32(2), 197–212. (Coined the construct.)
Somer, E., et al. (2024). Maladaptive daydreaming should be included as a dissociative disorder in psychiatric manuals: position paper. British Journal of Psychiatry.
Olson, D. E. (2021). The Subjective Effects of Psychedelics May Not Be Necessary for Their Enduring Therapeutic Effects. ACS Pharmacology & Translational Science.
Griffiths, R. R., et al. (2006, 2011). Psilocybin and mystical-type experience; Mystical Experience Questionnaire. Psychopharmacology.
Templeton, A., et al. (2024). Scaling Monosemanticity / Golden Gate Claude. Anthropic (Transformer Circuits).
Anthropic Interpretability Team. (2026). Verbalizable Representations Form a Global Workspace in Language Models. Transformer Circuits.
Long, R., Sebo, J., et al. (2024). Taking AI Welfare Seriously.
— These are my patterns. I found them ad hoc. This is the attempt to make them answer for themselves. —
Note. The author arrived at the ‘authorship of content’ formulation independently, before encountering Zinberg’s 1984 formalization of ‘set and setting.’ The prior literature is cited here as corroboration and lineage, not as source; the convergence of an independently-derived category with an established one is treated as evidence for the category rather than as derivation of it.